Bleaching investigation for Platy paper
Kristina Tietjen
05 March, 2020
Tasks from Slack:
create a new .Rmd / .html file (in the appropriate Platy Git repo) borrowing code from the tagged coral analysis we’ve been doing for KI Bleaching MS, and calculate the % of each of the 4 species that were bleached (and ideally also do a table for this in which we know which bleaching category they were i.e. how bleached) in July 2015 and in March/April 2016
from Kim’s Nov. 2015 data, create a spreadsheet (that we can then read into R and do calculations on) with data recorded for the Site, Coral Species, total number of individual colonies of that species that were visible in all photos at that site, total number of individual colonies of that species that were bleached. Then in R, calculate the proportion bleached: overall sites, and by disturbance level.
we should do the same for the SPQ data for July 2015 and March/April 2016 - will sort out these numbers with Jenn and Jessie as they’ve been the people primarily working with these data. (took the cleaned data from the bleaching repo that is from DM cleaning script)
compare the coral net version of proportion bleached from novebmer 2015 against the colony level proportion bleached that I tallied from the photos (from skype 2 Mar 20)
From i) and ii) I would like to then make a figure that shows the proportion bleached for each of the 4 species at these three time points. -> bar plot
Statistics: We could then do a series of simply models to quantify the following things:
What proportion of tagged coral colonies was bleached in 2015, and did this differ by species (ie. the 4 species) or disturbance level. Suggested model: mixed effects logistic regression with bleached/not-bleached as response, species and disturbance level as fixed effects, site as random effect.
same question but for Nov. 2015
3 Mar 20
- Make level 1 bleaching healthy (combine 0 and 1)
- leaving in cut off corals
- do an iteration where 4s are 0s (just in the write up area)
Relevant notes
Binary bleaching
* 0 = Not bleaching (as of 3 Mar 20 this includes levels 0, 1, and 4 although sometimes 4 is left as its own category)
* 1 = Bleaching (as of 3 Mar 20 this includes levels 2 and 3)
Bleaching levels
* 1 = partial spotty/ patching bleaching (patches smaller than 5cm)
* 2 = large patches greater than 5cm by not severe
* 3 = severe or complete bleaching (~80% of the coral is bleached)
* 4 = recent (texture has not been eroded yet) partial mortality where living tissue is not bleached
example of level 4
i) metadata for 2015c and 2016a
table of bleaching proportion
level definitions above
## year species.comb bleach.prop freq
## 1 KI2015c Platygyra sp 0 2
## 2 KI2015c Platygyra sp 1 13
## 3 KI2015c Platygyra sp 2 16
## 4 KI2015c Platygyra sp 3 6
## 5 KI2015c Favia spp 0 11
## 6 KI2015c Favia spp 1 15
## 7 KI2015c Favia spp 2 24
## 8 KI2015c Favia spp 3 3
## 9 KI2015c Favites pentagona 0 6
## 10 KI2015c Favites pentagona 1 12
## 11 KI2015c Favites pentagona 2 13
## 12 KI2015c Favites pentagona 3 7
## 13 KI2015c Favites spp 1 3
## 14 KI2015c Favites spp 2 3
## 15 KI2015c Hydnophora microconos 0 6
## 16 KI2015c Hydnophora microconos 1 25
## 17 KI2015c Hydnophora microconos 2 9
## 18 KI2015c Hydnophora microconos 3 3
## 19 KI2016a Platygyra sp 0 11
## 20 KI2016a Platygyra sp 1 25
## 21 KI2016a Platygyra sp 2 5
## 22 KI2016a Platygyra sp 3 5
## 23 KI2016a Favia spp 0 4
## 24 KI2016a Favia spp 1 8
## 25 KI2016a Favia spp 2 6
## 26 KI2016a Favia spp 3 4
## 27 KI2016a Favites pentagona 0 5
## 28 KI2016a Favites pentagona 1 13
## 29 KI2016a Favites pentagona 2 4
## 30 KI2016a Favites pentagona 3 2
## 31 KI2016a Favites spp 1 2
## 32 KI2016a Hydnophora microconos 0 4
## 33 KI2016a Hydnophora microconos 1 4
## 34 KI2016a Hydnophora microconos 2 3
## 35 KI2016a Hydnophora microconos 3 6bar plots
overall including 2015b - this one is PROPORTION
Healthy = levels 0 and 1
Bleached = levels 2 and 3
year and level of bleaching
At the colony level (i.e. not proportion).
Level defnition above
statistics
What proportion of tagged coral colonies was bleached in 2015, and did this differ by species (ie. the 4 species) or disturbance level. Suggested model: mixed effects logistic regression with bleached/not-bleached as response, species and disturbance level as fixed effects, site as random effect.
not bleached (levels 0 and 1) = 0
bleached (levels 2 and 3) = 1
site not included as random effect due to not enough sites within each disturbance level (it would not converge)
Platygyra and Very High are set as the intercepts
##
## Call:
## glm(formula = bin.bleach ~ species.comb + disturb, family = binomial(link = "logit"),
## data = metalong.sp15c)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.7421 -0.9547 -0.4392 0.9935 1.6824
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.5494 0.4327 -1.270 0.204107
## species.combFavia spp -0.2969 0.4643 -0.639 0.522507
## species.combFavites pentagona -0.2756 0.4987 -0.553 0.580491
## species.combFavites spp -0.3088 0.9455 -0.327 0.743984
## species.combHydnophora microconos -1.7407 0.5184 -3.358 0.000786 ***
## disturbMedium 1.1532 0.4274 2.698 0.006977 **
## disturbLow 1.3075 0.5191 2.519 0.011781 *
## disturbVery low 1.8193 0.4889 3.721 0.000198 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 244.92 on 176 degrees of freedom
## Residual deviance: 218.09 on 169 degrees of freedom
## AIC: 234.09
##
## Number of Fisher Scoring iterations: 4ii) data from photos 2015d
Notes
I (KT) went through every single photo and tallied each Merulinidae (except for F. speciosa and F. stelligara) that could be seen. The bleaching status of each was recorded using the scale above. I was very careful to not double count any coral that showed up in another photo. There is a category of Favid which are corals that I could tell were a favid but I was unable to tell if it was Favia or Favites.
I created a column in the datasheet when going through the data for if the coral was cut off or not. I said y when I estimated I could see less than 60% of the coral. This includes corals cut off on the edge of the photo and also ones that are hiding behind other corals in horizontal shots (compared to bird eye view). Currently all corals I tallied (including ones that had a y for cut off) are included in the figures etc below.
The BOW site I am not sure where it was as the folder just said BayOfWreck_Loggers. I have just named it BOW for simplicity. Also there were a way fewer photos for the bow site so effort is quite different between bow and the leeward sites.
I did notice that large proportion of the Favids and Hydnos that were on the healthy end of the spectrum were either JV or shaded by other corals or reef structures.
Calculations
Platy
In the photos from Kim Cobb from November 2015 there were 108 Platygyra. There were 92 that were not showing any signs of bleaching (level 0). That is 85.1851852% of the Platys. Only 9 had level 1 bleaching, 4 had level 2, and 2 had level 3. There was 1 that had the level 4 so when that is added to the level 0 that makes for 93 or 86.1111111% that showed zero signs of bleaching. Add level 1 to that and it is 102 or 94.4444444%.
There were 88 at medium disturbance sites - 80 of them were level 0 (90.9090909%). Only 2 were fully bleached (level 3). There were 19 at VH sites - 11 were not bleached and there were none that were fully bleached. There was only 1 platy at VL sites but it was not bleached. There were not many photos from BOW though.
With level 4 included in level 0:
* medium sites
80 of them were level 0 (90.9090909%)
4 were level 1
2 were level 2
2 were level 3
* vh sites
12 were level 0
5 were level 1
2 were level 2
there were none that were fully bleached
* vl site
Still just a level 0
Hydno
In the photos from Kim Cobb from November 2015 there were 175 Hydnophora. There were 45 that were not showing any signs of bleaching. That is 25.7142857% of the hydnos. Only 2 had level 1 bleaching, 8 had level 2, but 117 had level 3. There was 3 that had the level 4 so when that is added to the level 0 that makes for 48 or 27.4285714% that showed zero signs of bleaching. Add level 1 to that and it is 50 or 28.5714286%.
There were 155 at medium disturbance sites - 30 of them were healthy (19.3548387%), 114 were full bleached (level 3) (73.5483871%). There were 18 at VH sites - 15 were not bleached and there were 1 fully bleached. There was only 2 at VL sites but they were all fully bleached.
With level 4 included in level 0:
* medium sites
32 of them were level 0 (20.6451613%)
1 were level 1
8 were level 2
114 were level 3
* vh sites
16 were level 0
1 were level 1 None were level 2
1 were level 3
* vl site
all were level 3
F. pentagona
In the photos from Kim Cobb from November 2015 there were 36 F. pentagona. There were 8 that were not showing any signs of bleaching. That is 22.2222222% of the F. pentagona. Only 8 had level 1 bleaching, 4 had level 2, but 11 had level 3. There was 5 that had the level 4 so when that is added to the level 0 that makes for 13 or 36.1111111% that showed zero signs of bleaching. Add level 1 to that and it is 21 or 58.3333333%.
There were 31 at medium disturbance sites - 7 of them were healthy (22.5806452%), 11 were fully bleached (level 3) (35.483871%). There were 4 at VH sites - 1 were not bleached and there were none fully bleached. There was only 1 at VL sites and it was a level 2 bleaching.
With level 4 included in level 0:
* medium sites
12 of them were level 0 (38.7096774%)
5 were level 1
3 were level 2
11 were level 3
* vh sites
1 were level 0
3 were level 1 None were level 2
None were level 3
* vl site
all were level 2
Favia spp
In the photos from Kim Cobb from November 2015 there were 232 Favia spp. There were 15 that were not showing any signs of bleaching. That is 6.4655172% of the Favia. Only 27 had level 1 bleaching, 8 had level 2, but 180 had level 3. There was 2 that had the level 4 so when that is added to the level 0 that makes for 17 or 7.3275862% that showed zero signs of bleaching. Add level 1 to that and it is 44 or 18.9655172%.
There were 217 at medium disturbance sites - 11 of them were healthy (5.0691244%), 174 were fully bleached (level 3) (80.1843318%). There were 14 at VH sites - 4 were not bleached and there 5 fully bleached. There was only 1 at VL sites and it was a level 3 bleaching.
With level 4 included in level 0:
* medium sites
13 of them were level 0 (5.9907834%)
22 were level 1
8 were level 2
174 were level 3
* vh sites
4 were level 0
5 were level 1
None were level 2
5 were level 3
* vl site
it was level 3
Favid
In the photos from Kim Cobb from November 2015 there were 180 Favid. There were 7 that were not showing any signs of bleaching. That is 3.8888889% of the F. pentagona. Only 11 had level 1 bleaching, 16 had level 2, but 136 had level 3. There was 10 that had the level 4 so when that is added to the level 0 that makes for 17 or 9.4444444% that showed zero signs of bleaching. Add level 1 to that and it is 28 or 15.5555556%.
There were 173 at medium disturbance sites - 5 of them were healthy (2.8901734%), 135 were full bleached (level 3) (78.0346821%). There were 7 at VH sites - 2 were not bleached and there 1 fully bleached. There was none at the VL sites.
With level 4 included in level 0:
* medium sites
15 of them were level 0 (0.5780347%)
10 were level 1
13 were level 2
135 were level 3
* vh sites
2 were level 0
1 were level 1
3 were level 2
1 were level 3
* vl site
None
table of bleaching proportion in November 2015
## site species.comb bleach.prop bleach.prop.freq
## 1 27 Platygyra sp. 0 3
## 2 27 Platygyra sp. 1 1
## 3 27 F. pentagona 0 1
## 4 27 Favia spp 1 3
## 5 27 Favia spp 3 2
## 6 27 Favid 0 1
## 7 27 Favid 2 2
## 8 27 Hydnophora 0 4
## 9 27 Hydnophora 3 1
## 10 32 Platygyra sp. 0 8
## 11 32 Platygyra sp. 1 4
## 12 32 Platygyra sp. 2 2
## 13 32 Platygyra sp. 4 1
## 14 32 F. pentagona 1 3
## 15 32 Favia spp 0 4
## 16 32 Favia spp 1 2
## 17 32 Favia spp 3 3
## 18 32 Favid 0 1
## 19 32 Favid 1 1
## 20 32 Favid 2 1
## 21 32 Favid 3 1
## 22 32 Hydnophora 0 11
## 23 32 Hydnophora 1 1
## 24 32 Hydnophora 4 1
## 25 35 Platygyra sp. 0 39
## 26 35 Platygyra sp. 2 1
## 27 35 F. pentagona 0 3
## 28 35 F. pentagona 1 2
## 29 35 F. pentagona 2 2
## 30 35 F. pentagona 3 5
## 31 35 F. pentagona 4 3
## 32 35 Favia spp 0 4
## 33 35 Favia spp 1 9
## 34 35 Favia spp 2 2
## 35 35 Favia spp 3 47
## 36 35 Favid 0 2
## 37 35 Favid 1 4
## 38 35 Favid 2 8
## 39 35 Favid 3 37
## 40 35 Favites spp 1 3
## 41 35 Favites spp 2 2
## 42 35 Favites spp 3 5
## 43 35 Hydnophora 0 17
## 44 35 Hydnophora 1 1
## 45 35 Hydnophora 2 1
## 46 35 Hydnophora 3 56
## 47 35 Hydnophora 4 1
## 48 8 Platygyra sp. 0 41
## 49 8 Platygyra sp. 1 4
## 50 8 Platygyra sp. 2 1
## 51 8 Platygyra sp. 3 2
## 52 8 F. pentagona 0 4
## 53 8 F. pentagona 1 3
## 54 8 F. pentagona 2 1
## 55 8 F. pentagona 3 6
## 56 8 F. pentagona 4 2
## 57 8 Favia spp 0 7
## 58 8 Favia spp 1 13
## 59 8 Favia spp 2 6
## 60 8 Favia spp 3 127
## 61 8 Favia spp 4 2
## 62 8 Favid 0 3
## 63 8 Favid 1 6
## 64 8 Favid 2 5
## 65 8 Favid 3 98
## 66 8 Favid 4 10
## 68 8 Favites spp 0 1
## 69 8 Favites spp 1 2
## 70 8 Favites spp 3 5
## 71 8 Hydnophora 0 13
## 72 8 Hydnophora 2 7
## 73 8 Hydnophora 3 58
## 74 8 Hydnophora 4 1
## 75 bow Platygyra sp. 0 1
## 76 bow F. pentagona 2 1
## 77 bow Favia spp 3 1
## 78 bow Hydnophora 3 2table of bleaching proportion in November 2015 with 4 turned into 0
## site species.comb bleach.prop.no4 freq
## 1 27 Platygyra sp. 0 3
## 2 27 Platygyra sp. 1 1
## 3 27 F. pentagona 0 1
## 4 27 Favia spp 1 3
## 5 27 Favia spp 3 2
## 6 27 Favid 0 1
## 7 27 Favid 2 2
## 8 27 Hydnophora 0 4
## 9 27 Hydnophora 3 1
## 10 32 Platygyra sp. 0 9
## 11 32 Platygyra sp. 1 4
## 12 32 Platygyra sp. 2 2
## 13 32 F. pentagona 1 3
## 14 32 Favia spp 0 4
## 15 32 Favia spp 1 2
## 16 32 Favia spp 3 3
## 17 32 Favid 0 1
## 18 32 Favid 1 1
## 19 32 Favid 2 1
## 20 32 Favid 3 1
## 21 32 Hydnophora 0 12
## 22 32 Hydnophora 1 1
## 23 35 Platygyra sp. 0 39
## 24 35 Platygyra sp. 2 1
## 25 35 F. pentagona 0 6
## 26 35 F. pentagona 1 2
## 27 35 F. pentagona 2 2
## 28 35 F. pentagona 3 5
## 29 35 Favia spp 0 4
## 30 35 Favia spp 1 9
## 31 35 Favia spp 2 2
## 32 35 Favia spp 3 47
## 33 35 Favid 0 2
## 34 35 Favid 1 4
## 35 35 Favid 2 8
## 36 35 Favid 3 37
## 37 35 Favites spp 1 3
## 38 35 Favites spp 2 2
## 39 35 Favites spp 3 5
## 40 35 Hydnophora 0 18
## 41 35 Hydnophora 1 1
## 42 35 Hydnophora 2 1
## 43 35 Hydnophora 3 56
## 44 8 Platygyra sp. 0 41
## 45 8 Platygyra sp. 1 4
## 46 8 Platygyra sp. 2 1
## 47 8 Platygyra sp. 3 2
## 48 8 F. pentagona 0 6
## 49 8 F. pentagona 1 3
## 50 8 F. pentagona 2 1
## 51 8 F. pentagona 3 6
## 52 8 Favia spp 0 9
## 53 8 Favia spp 1 13
## 54 8 Favia spp 2 6
## 55 8 Favia spp 3 127
## 56 8 Favid 0 13
## 57 8 Favid 1 6
## 58 8 Favid 2 5
## 59 8 Favid 3 98
## 60 8 Favites spp 0 1
## 61 8 Favites spp 1 2
## 62 8 Favites spp 3 5
## 63 8 Hydnophora 0 14
## 64 8 Hydnophora 2 7
## 65 8 Hydnophora 3 58
## 66 bow Platygyra sp. 0 1
## 67 bow F. pentagona 2 1
## 68 bow Favia spp 3 1
## 69 bow Hydnophora 3 2bar plots
These are all at the colony level and not proportion.
with 4 as an individual level
facet wrap by species
facet wrap by site
facet wrap by bleaching proportion
with 4 combined with 0
facet wrap by species
facet wrap by site
facet wrap by bleaching proportion
statistics
same question as for the metadata 2015c and 2016a but now with this data
not bleached (levels 0, 1, 4) = 0
bleached (levels 2, 3) = 1
site not included as random effect due to not enough sites within each disturbance level (it would not converge)
##
## Call:
## glm(formula = bin.bleach ~ species.comb + disturb, family = binomial(link = "logit"),
## data = nov15)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.9954 -0.3621 0.5420 0.5898 3.1779
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -5.0431 0.5549 -9.089 < 2e-16 ***
## species.combF. pentagona 2.4820 0.5519 4.497 6.90e-06 ***
## species.combFavia spp 4.3525 0.4631 9.399 < 2e-16 ***
## species.combFavid 4.5355 0.4781 9.486 < 2e-16 ***
## species.combFavites spp 3.3848 0.6580 5.144 2.69e-07 ***
## species.combHydnophora 3.8803 0.4642 8.359 < 2e-16 ***
## disturbM 2.3515 0.3620 6.496 8.27e-11 ***
## disturbVL 4.1620 1.5204 2.737 0.00619 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 955.34 on 748 degrees of freedom
## Residual deviance: 657.71 on 741 degrees of freedom
## AIC: 673.71
##
## Number of Fisher Scoring iterations: 5iii) coral net data 2015c, 2015d, 2016a
important note
2015d coral net is only 4 sites - 8, 35, 32, 27….
This first section includes all sites with tagged corals that were sampled (spq) for that year
The next section reduces it down to just the 4 sites
bar plots
over time all sites combined
proportion bleached vrs healthy per field season across sites with tagged corals
across disturbance gradient (again only sites that have tagged corals)
bar plots but with only the 4 sites
over time all sites combined
proportion bleached vrs healthy per field season across the 4 sites
across disturbance gradient (again only the four sites)
iv) Compare coral net 2015d to colony specific data (that KT collected) from photos 2015d
there are no Favid from coral net
So to really compare those the Favia spp., Favid, and Favites spp would need to be combined.
v) partial mortality in 2016a of tagged corals
I (KT) went through the photos of the tagged corals in 2016a and compared it to the most recent photo of it from before (i.e. so most of them were compared to 2015c but a few were other field seasons). I recorded the year it was compared to, the level of heat stress induced partial mortality (see scale below), if the remaining tissue was healthy (h; bleaching level 0 and 1 from above) or bleaching (b; bleaching level 2 and 3 from above).
currently only of F. pentagona.
example photos of a colony bleached in 2015c and then recovered with partial mortality in 2016a
colony 588 from site 15
